Nothomyrmecia macrops: a living-fossil ant rediscovered.

Reprint Series

15 September 1978, Volume 201, pp. 979-985

o
Nothomyrmecia macrfos:

A Living-Fossil Ant Rediscovered

Robert W. Taylor

46

Copyright 1978 by the American Association for the Advancement of Science

Nothomyrmecia macrops:
A Living-Fossil Ant Rediscovered

The most primitive living ant, previously an enigma,
rediscovered and the subject of international study.

Robert W. Taylor

The Australian ant Nothomyrmecia
macrops was described by Clark in 1934
(/) from two worker specimens in the
National Museum of Victoria, Mel-
bourne. The species was classified in a
new monotypic higher taxon, now tribe
NothomyrmSciini of subfamily Myrme-
ciinae (1-3). Other myrmeciine genera
are Myrmecia (Australia, about 65 spe-
cies; New Caledonia, 1 species), Priono-
myrmex (Oligocene, Baltic Amber, 1
species), andAmeghinoa (early Tertiary,
Argentina, 1 species) (4). The myrme-
ciines are considered the most structur-
ally generalized of all ants, apart from
the North American Cretaceous fossil
Sphecomyrma freyi (subfamily Spheco-
myrminae) (5). Myrmecia , while fully eu-

SCIENCE, VOL. 201, 15 SEPTEMBER 1978

social, has significantly primitive behav-
ior (6, 7). Nothomyrmecia has been rec-
ognized as the most generalized of these
insects and hence the most primitive
known living ant, the descendent of a
group important in formicid phylogeny
(5, 7), and a likely near facsimile of spe-
cies extant perhaps 60 million years ago
or more. There has been speculation on
the outside possibility that its behavior,
when known, might represent an early
stage in formicid social evolution (3, 7,
8). Study of the Melbourne specimens
has been limited by their being dry-
mounted, while the developmental
stages and adult sexual forms have re-
mained unknown. In the absence of fur-
ther collections, N. macrops has be-

come, naturally enough, a "holy grail"
to ant specialists, and its "rediscovery in
the living condition" has been stated as
"one of the principal challenges of mod-
ern Australian entomology" (8).

Clark's specimens were probably col-
lected near the western end of the Great
Australian Bight by an excursion party
that traveled, in December 1931, south-
ward from near Balladonia through mal-
lee-type Eucalyptus woodland and for-
est and set up camp for several weeks at
the Thomas River mouth, east of Espe-
rance, in the extensive sand plain heath
present there. Insects were collected,
without precise data, for a local natural-
ist, Mrs. A. E. Crocker, who sent them
to Clark. Many Australian and American
collectors and expeditions have since un-
successfully sought Nothomyrmecia in
this area, especially in the sand plain
heath, where a guild of similarly pale col-
ored, large-eyed, nocturnally foraging
ants is well represented (8).

Nothomyrmecia was rediscovered on
22 October 1977 southeast of Ceduna on
the Eyre Peninsula of South Australia by
a CSIRO field party that had camped
overnight en route from Canberra to
Western Australia (9). Workers and de-
alate queens were collected while forag-
ing nocturnally on the ground and tree
trunks in disturbed roadside mallee
woodland, but colonies were not lo-

The author is a Principal Research Scientist and
Curator of Hymenoptera (Formicidae) with the Aus-
tralian National Insect Collection. Entomology Divi-
sion, CSIRO, Canberra.

0036-8075/78/0915-0979$01.50/0 Copyright 1978 AAAS

979

(that is, jugal) lobe is present on the hind
wing, as is the case in various primitive
ants; but a series of basal hamuli in addi-
tion to the apical set occurs only in Noth-
omyrmecia among formicids and is a
primitive character (/5). The forelegs of
most, but not every, male specimen each
bear two apical tibial spurs, one a long
calcar, the other a short thick cone. All
other Hymenoptera, except some primi-
tive Symphyta and Proctotrupoidea and
all Ceraphronoidea, have single fore-
tibial spurs (18). Some illustrations of
Sphecomyrma freyi (5) incorrectly depict
paired spurs in that species. The me-
soscutum carries parapsidal furrows, but
notauli are lacking, a clearly derived
condition. A functional ventral stridula-
tory organ, like that of the female castes,
is present. Significance of genitalic struc-
tures cannot be estimated in the absence
of comparative studies. Cerci are pres-
ent; sternite IX has a bifid apex; the
gonoforceps are divided into proximal
and distal sections and lack the dorsal
projection present in some Myrmecia
species (19); the volsellae are unspecial-
ized, and the penis valves are strikingly
elaborate, with several denticulate apical
processes.

Adult Nothomyrmecia and Myrmecia
share other significant primitive features
including a formula of six maxillary and
four labial palpae in all sexes; 12 anten-
nal segments in female castes and 13 in
males; paired calcariae on the middle
and hind tibiae; tarsal claws each having
a strong median tooth; and a sting whose
complete structure includes a furcula
and two-jointed gonostyli, as in other
primitive ants (20). The proventriculus of
workers is actively dammed, with the
cuticular structure relatively unspecial-
ized, that of Nothomyrmecia being simi-
lar to Pseudomyrmex (Pseudomyr-
mecinae) (21). Most of these features are
present in various other primitive ants
of subfamilies Ponerinae and Pseudo-
myrmecinae.

Nothomyrmecia eggs, like those of
Myrmecia (6) are subspherical and non-
adhesive. The larvae lack specialized tu-
bercles and have a primitive shape and
general structure, sharing many features
with Myrmecia (22) and with primitive
prodoryline and proponerine Ponerinae
(23), although the sensilla on the mouth-
parts are more abundant. The pupal co-
coons are substantial, but have thinner
walls than those of Myrmecia and most
primitive ponerines; meconia are pro-
duced.

In short, the general structure of
adults and developmental stages con-
firms the primitiveness of Nothomyr-
mecia. Clearly derived features include

15 SEPTEMBER 1978

Fig. 2. Nothomyrmecia wings. (A) Male; (B)
outline of female wing to same scale; and (C)
detail of enlarged female wing to illustrate
characteristics of its structural reduction.
Scale bars, 2 mm.

the vestigial worker ocelli, female bra-
chyptery, and male mesoscutal struc-
ture, whereas the abdominal and female
mandibular structure, with the basal
hamuli and two-segmented gonoforceps
of the males, are features more primitive
than those of Myrmecia. The stridula-
tory organ and supplementary male fore-
tarsal- spurs probably represent evolu-
tionary novelties.

The differences in waist structure
separating Nothomyrmecia and the ad-
vanced myrmeciines are neither superfi-
cial nor trivial. Abdominal segment IV of
Myrmecia (Fig. 3B) is "tubulate" in

form, with its acrotergum and acroster-
num each broadly expanded and sepa-
rated from the body of its plate
by a transverse construction and incised
groove, presumably representing the an-
tecostal suture (24). In addition, the lat-
eral edges of these plates are aligned and
intimately associated, especially in the
anterior portion, which at rest is fully in-
serted into segment III (the postpetiole),
the posterior edge of which embraces the
antecostal constriction of segment IV to
form a ball joint. Controlled telescopic
and rotational movement at this node fa-
cilitates abdominal mobility. Similar tu-
bulation of segment III was doubtless
important in evolution of the petiole-
postpetiole joint in ants and other Hyme-
noptera. Evolution of the mesosomal-
metasomal waist in primitive Aprocrita
apparently involved acrotergal expan-
sion (25). Subfamily Pseudomyrmecinae
has been related to the Myrmeciinae (?)
and has a Myrmecia-like metasoma. Tu-
bulation of abdominal IV is found in
workers, queens, and most males of sub-
family Ponerinae (Fig. 3 A), where full
lateral fusion of the sclerites is frequent
and the structure is sometimes further
elaborated (26). Tubulation evidently
functions as a preadaptation to post-
petiole formation, which has been impor-
tant in evolution within and beyond the
Ponerinae (see below and 27).

In contrast, Nothomyrmecia (Fig. 3C)
has the sclerites of segment IV freely ar-
ticulated, with structure similar to that of
succeeding segments. The acrotergite
and acrosternite are probably represent-
ed only by the thickened anterior rims of
the sclerites, and the portion of each
plate which inserts into segment HI is

Fig. 3. Diagrams of ex-
ploded abdominal
plates of primitive ants,
illustrating anatomical
differences in segment
IV, which has its scler-
ites separate in Noth-
omyrmecia (C), but as-
sociated to form a tubu-
late structure in Myr-
mecia (B) and Ambly-
opone (A).

MYRMECIA

,^ rv\

981

only feebly differentiated, by superficial
sculpture (which includes the sternal
stridulitrum) from the exposed part. The
structure is here nontubulate, and a post-
petiole is not differentiated. A generally
Nothomyrmecia-\\ke segment IV is con-
sidered primitive to that of Myrmecia,
because this is the condition in most acu-
leates, and probably in Sphecomyrma
(5).

Phylogenetic Implications

Current evolutionary models (3, 5, 7)
recognize two ant phylads, the myrme-
ciioid and poneroid complexes. Noth-
omyrmecia represents primitive Myrme-
ciinae, which stand near the base of the
myrmeciioid complex, linking the ances-
tral Sphecomyrminae and subfamily
Aneuretinae (28), which apparently gen-
erated the major subfamilies Dolichode-
rinae and Formicinae. The advanced
myrmeciines (Myrmecia, Prionomyrmex,
Ameghinoa) and Pseudomyrmecinae are
derived separately from Nothomyr-
mecia -grade stock. The Poneroid phylad
includes the five other ant subfamilies,
which are derived through the primitive
Ponerinae, as possibly represented by
Amblyopone (tribe Ambyloponini).

As indicated above, the advanced
myrmeciines and the pseudomyrmecines
have abdominal IV tubulate, like the
Ponerinae. Also, the pseudomyrfne-
cihes, like ponerines and unlike Noth-
omyrmecia or Myrmecia, have dorsal
stridulatory organs, which are found no-
where else in the myrmeciioid complex

MYRMICINAE

DORYLINAE, ECITONINAE,
LEPTANILLINAE

PSEUDOMYRMECINAE

(14). If the structures concerned are truly
homologous (that is, each uniquely
evolved), the pseudomyrmecines and ad-
vanced myrmeciines must be dissociated
from Nothomyrmecia by transfer to the
poneroid complex, with derivation from
a primitive poneroid ancestor shared
with the Ponerinae and possessing a
tubulate abdominal IV bearing a dorsal
stridulatory organ. Loss of such an organ
in the Myrmecia lineage could easily
have occurred it has within several
ponerine and myrmicine genera (14).
This hypothesis is summarized and ex-
tended in Fig. 4. Nothomyrmecia re-
mains approximately annectent to the
Aneuretinae, and the model requires
reinstatement for it of Clark's subfamily
Nothomyrmeciinae (2), which is prob-
ably desirable anyway on phenetic
grounds. Also, the erstwhile "myrme-
ciioid complex" would need renaming,
as the "formicoid complex." This model
collapses if ponerine-like tubulation and
dorsal stridulatory organs arose con-
vergently with the higher myrmeciines
and pseudomyrmecines from a primi-
tive Nothomyrmecia-Te\a.ted stock. The
many similarities between Nothomyr-
mecia and Myrmecia do not invalidate
the hypothesis. These ants would be ex-
pected to retain primitive characteristics
in common, even if exemplifying dif-
ferent basal lineages with separate evolu-
tionary potential which we, in hindsight,
see realized among modern ants.

Formicoid trophic evolution has
largely involved adaptations to liquid
feeding by adults and larvae. These in-
clude convergent development, in both

FORMICINAE

DOUCHODERINAE

ANEURETINAE

NOTHOMYRMECIINAE Aneurelus

Nothomvrmeciq

tubulotion of'
obdominal
segment TZ

PONEROID COMPLEX

ant-like mandibles
and antennae
SPHECOMYRMINAE

true sociality

metapleural
gland

FORMICOID COMPLEX

ANCESTOR

Fig. 4. A simple, hypothetical, branching phylogenetic diagram, illustrating the hypothesis de-
veloped here. Nothomyrmecia stands close to the ancestors of the formicoid phylad, and Myr-
mecia to the poneroid stock.

982

Dolichoderinae and Formicinae, of elab-
orate proventriculi, to serve as passive
valve dams, retaining liquid food in the
greatly expandable crop (21, 29). This is
often accompanied by gastral expansion,
facilitated partly by free movement of
the segment IV sclerites. These trends
culminate in the separate development of
replete ("honey pot") workers in several
groups of both subfamilies (7). These are
capable of enormous gastral expansion,
and function as stationary liquid storage
reservoirs for their colonies. The begin-
nings of this evolution could be repre-
sented by the crop and gastral expansion
observed in Nothomyrmecia foragers
(see below). Correlated trends include
reduction in gastral sclerotization and
convergent modification of sting struc-
ture and function (to either a spreading
or a spraying device) in the two sub-
families.

These developments might have been
forced by problems of mechanical op-
eration of a piercing sting occasioned
by crop expansion. No known formi-
coid except Nothomyrmecia has an
abdominal stridulatory organ (14). The
"primitive formicoids" of Fig. 4 need
not have had such a structure; but if they
did, the organ (whether dorsal or ventral)
could have been lost in further evolution,
in correlation with sclerotic reduction of
the gaster. Also, its operation would
likely have been compromised in groups
with gastral expansion exceeding that of
Nothomyrmecia. The Sphecomyrma fos-
sils have been reported to lack a dorsal
stridulatory organ (14), but they should
be checked for the presence or absence
of a ventral one. Abdominal tubulation is
a preadaptation to postpetiolar develop-
ment through strangulation of segment
IV. Among ants this occurs only in the
poneroid complex (as comprised here)
and has occurred repeatedly (27). The
only other known postpetiolate Hyme-
noptera (Bradynobaenidae-Apteryogy-
ninae) (/5) apparently also have ab-
dominal IV tubulate (30). Tubulation
aids controlled use of the sting, which is
rarely degenerate in poneroids; it must,
however, limit crop expansion and has
perhaps restricted emphasis on liquid
feeding in the complex (29). Repletes are
unknown among poneroids; the proven-
triculus is seldom elaborated, and is re-
duced to a simple tube in some groups
(21). The myrmicines especially have de-
veloped alternative trophic life-styles,
utilizing nonliquid foods such as fungi,
seeds, and plant material (7). Massive
gastral expansion, comparable to that of
formicoid repletes, occurs in several po-
neroid groups having physogastric
queens, whose abdomens swell to an un-

SCIENCE, VOL. 201

usual degree because of hypertrophy of
the ovaries and fat body. In the minute
European workerless parasite Tele-
utomyrmex schneideri (Myrmicinae),
segment IV is not involved in this expan-
sion (31). Queens of the various army
ants Aenictus (Dorylinae), Eciton,
Labidus, and Neivamyrmex (Ecitoninae),
and probably those of Leptanillinae,
oviposit intermittently and massively.
Physogastry here does involve expan-
sion of segment IV, which has separable
sclerites; and these females, unlike their
workers, have a single waist node (32).
This apparently reversed evolutionary
trend is doubtless an adaptation to phy-
sogastry. It neatly correlates free move-
ment of the segment IV sclerites nega-
tively with the presence of postpetioles
and positively with gastral expansion,
and explains the peculiar female meta-
somal dimorphism in these ants. Com-
parable reversal of gastral tubulation is
seen in army ant males and in those of
the advanced (euponerine) Ponerinae.

Many of the major characteristics of
ant phylogeny (Fig. 4) which are re-
viewed above can thus be correlated
with the structural modifications of ab-
dominal segment IV which have pro-
foundly influenced ant evolution.

Genetics

Preliminary investigations (33) show
that Nothomyrmecia has a very high
diploid chromosome number of about 92,
the highest number known among
Hymenoptera, where the range other-
wise is 2n = 6 to 84. This range is en-
compassed by the ants, and all Hyme-
noptera with In exceeding 52 are ants.
Most are Myrmecia species, of which ten
have been investigated yielding a In
range of 9 to 84. Four species of Myr-
mecia, with 2n = 60, 66, 81, and 84,
have diploid numbers exceeding 52, and
a Bothroponera species (Ponerinae) has
2n = 60 (34). Most Nothomyrmecia
chromosomes are dotlike acrocentrics,
but some larger metacentrics are pres-
ent.

Preliminary electrophoretic study (35)
of 18 loci, in about 100 foraging workers,
has revealed one polymorphic locus
(amylase) showing substantial variation
(four alleles at frequencies greater than
.05). This suggests either a greater ef-
fective population size than field obser-
vations indicate or stabilizing selection
at this or a closely linked locus. Marker
genes at the amylase locus should pro-
vide information on the multiplicity of
insemination of queens and the parent-
age of males.

15 SEPTEMBER 1978

Field Studies

The woodland at the Eyre Peninsula
collection site seems similar to that in
surrounding areas. Eucalyptus oleosa
(Myrtaceae) forms an almost pure stand
on fine textured, well drained, brown
calcareous .earth with low organic con-
tent. The presence of scattered Callitris
preissii (Cupressaceae), which forms an
almost pure stand nearby and is else-
where very patchily distributed, might
indicate soil or drainage peculiarities.
The apparent absence of the otherwise
locally common funnel ant Aphaeno-
gaster barbigula (Myrmicinae), which
builds easily collapsed, craterlike, pitfall
trapnest entrances up to about 25 cen-
timeters in diameter, might be related to
soil type, and could be important in al-
lowing access to the site by Nothomyr-
mecia. The plants present (36) are all
widespread species found on both sides
of the Bight; most range eastward into
Victorian mallee woodland.

Nothomyrmecia workers were readily
collected on Eucalyptus trunks at night,
but they could not be found during day-
time. Night search in surrounding areas
shows that the population is apparently
very local, occupying only several hect-
ares. The ant seems absent from appar-
ently similar adjacent sites. Nest en-
trances, when located, consisted of small
(about 4 to 6 mm), unspecific holes in the
ground under shallow leaf litter, without
surrounding mounds or deposited soil.
Location of further populations will al-
most certainly require night search,
since the presence of Nothomyrmecia is
not evidenced during the day. The ants
forage singly and range to the tops of the
trees, where they probably seek sweet
substances and hunt for small arthro-
pods. The remains of a small uniden-
tifiable microlepidopteran and a spider-
ling have been taken from foragers.
Workers feed avidly from baits of honey
streaked on the tree trunks Su^S imbibing
for long periods, often exceeding 30 min-
utes. Their gasters visibly expand with
crop expansion, although the sclerites do
not separate sufficiently to expose the in-
tersegmental membranes. After feeding,
foragers continue to stray on the trees,
apparently randomly, and observers ex-
perience great difficulty in tracking the
return of the ants to nests, at least until
around dawn. At first light, the ants be-
gin to leave the trees, proceeding direct-
ly and positively across the surrounding
leaf litter to the nest entrances, which
are located in open ground and are evi-
dently not associated with basal accumu-
lations of bark litter and debris near
trees. Incomplete observations, inter-

rupted by bad weather, suggest that
there might be a considerable exodus
from the nests at dusk, with few foragers
returning until near dawn. If this is true,
the position of the sun or light areas of
the sky would be similarly related to the
body axis on both journeys. There is no
evidence that chemical trails are laid by
foragers, and, unlike smaller Myrmecia
species, they show no structured jump-
ing or hopping behavior. Disturbed for-
agers sometimes adopt a stationary
open-jawed threatening stance, but they
usually fall abruptly to the ground and
feign death in a cryptic, motionless pupal
posture. A Myrmecia species, similar in
size to Nothomyrmecia , is an equivalent
diurnal forager, first appearing as Noth-
omyrmecia withdraws. A few dealate
queens of Nothomyrmecia were encoun-
tered among the workers. When tracked
to nests they proved to be colony found-
ers, following the incompletely claustral
mode of establishment, in which young
queens leave their immature first brood
in order to forage.

Bionomics

Five Nothomyrmecia nests were ex-
cavated on 17 to 18 November. In each a
single gallery (diameter, 4 to 5 mm) de-
scended steeply at about 60 to a termi-
nal, subelliptical, horizontal chamber
(diameter, 3 to 5 cm; height, 5 to 10 mm;
depth below ground, ranging from 18 to
43 cm). Each shaft had three to five side
chambers, one within 10 cm of the sur-
face. The ants retreated timidly, most
being captured with brood and queens
(one in each nest) in the terminal galler-
ies. The brood comprised numerous half-
to full-sized larvae, all probably of a
single generation, and a few pupae; eggs
were not seen. Callow adults, alate
queens, and males were absent. Mature
nests probably contain 50 to 70 workers.
A dense layer of calcrete rocks in the soil
profile' apparently limited nest depth. A
sixth nest penetrated this, and excava-
tion was abandoned.

By 9 March 1978, four colonies, two
with queens, survived in laboratory cul-
ture. Additional larvae had not been pro-
duced, and all but a few in two nests had
pupated. Workers began emerging from
24 December, males from 16 January,
queens from 20 January. Callows are
recognizable for about 2 days. Emer-
gence seemed due for completion by
mid- April. One colony produced males,
and another males plus queens. The lat-
ter began emerging on 20 January and 3
March, respectively, possibly in a natu-
ral sequence. Both trophic and apparent-

983

ly reproductive eggs were laid occasion-
ally, but consistently, by workers. These
were fed mostly to larvae, but also to
other adults, including mother queens
and alate sexuals. Oviposition by queens
was not observed. No reproductive eggs
were accumulated. Two observation se-
ries of assembled foragers were collected
on 23 October and 17 to 18 November.
The first included a (presumably colony-
founding) dealate female. Both of these
groups accumulated many reproductive
eggs from late December onward, and a
few of these hatched by mid-February;
pupation had not occurred by 9 March. It
is not known whether the dealate female
in the mixed group contributed eggs, but
oviposition by workers was seen. Pro-
duction of eggs is thus not inhibited by
laboratory conditions. These events sug-
gest that brood is normally not present in
colonies during winter (say from late
April to early September). As a result,
foraging in this season is likely to be re-
duced, and nests possibly sealed, as in
Myrmecia tarsata , which does not over-
winter brood (personal observation).

A further nest, with a single terminal
chamber 18 cm below ground, was also
excavated. This contained two dealate
queens. Brood was not seen and, if pres-
ent, must have been small, this nest was
located by tracking a foraging queen at
dawn from a tree 4.5 meters away. One
queen left the nest briefly at night in light
rain to drink water from fallen leaves.
These ants survived in culture without
evident antagonism, and on 9 March
were supporting a large larva and a co-
coon, which was smaller than those in
mature colonies and was spun on 28 Feb-
ruary. Larvae were fed with insect frag-
ments and apparently normal (reproduc-
tive as opposed to trophic) eggs by both
queens. Founding thus may be pleome-
trotic with queens evidently reduced later
to one (secondary monogyny). Mating
flight details are unknown. The queens
might flutter from vegetation, like some
brachypterous Myrmecia (personal ob-
servation). None had undergone deala-
tion as virgins in the nests by 9 March.
Alates are presumably released by late
summer or autumn (March or April) but
might be overwintered in parent nests.
Founding queens excavate to consid-
erable depth; and, even if released in late
summer, evidently mature no eggs until
spring.

Aptery or brachyptery in Myrmecia
queens is not uncommon (2, 16, 37); al-
though its adaptiveness is unclear, it
might sometimes involve premating iso-
lating mechanisms. This seems unlikely
in Nothomyrmecia, where brachyptery
might relate to population structure, as

984

an adaptation developed in small scat-
tered populations held in enclaves by
competition with other ants, or by pre-
cise, unusual, ecological requirements.
It might be inadaptative for queens to
disperse and attempt to establish colo-
nies away from enclaves. The situation
could be compared to that of brachyp-
terous mountaintop or island insects
(38), and Nothomyrmecia could be dan-
gerously overspecialized in this regard.
Brachyptery might be recently evolved;
female wings, if nonfunctional for dis-
persal, would probably quickly dis-
appear altogether. Short-winged queens
might be irregular products of drought-
stressed colonies, as has been reported
in some Chelaner (Myrmicinae) species
in semiarid Australia (39, 40), but no
available evidence suggests such dimor-
phism in Nothomyrmecia.

Adult N. macrops are largely nectar-
ivorous but drink hemolymph from in-
sect prey. The latter, with little dis-
section, is fed directly to larvae. Larvae
can move independently toward food,
and cannibalism among them seems rare.
Pupae are used for larval food if forage is
withheld. Eggs are scattered in observa-
tion nests, with larvae plus eggs and pu-
pal cocoons only roughly segregated. Ma-
ture larvae swell anteriorly before spin-
ning cocoons and are buried by workers
to facilitate cocoon formation. Emer-
gence from cocoons is often assisted by
nurses, which tend to be the smaller,
least aggressive workers. Occasional
trophallaxis has been seen between
workers, and with sexuals or larvae,
which exude anal droplets imbibed by
workers. Workers and females actively
groom each other, with special attention
to the posterior mesosoma (? metapleural
glands). They will collapse tonically im-
mobile in pupal posture if nests are jolt-
ed, or if dragged or carried by nest mates.
Queens or workers may be dragged
by antennae or limbs without tonic
immobility, sometimes by counteracting
workers, and appear from abdominal
movements to stridulate if distressed by
this. Stridulation occurs when workers,
queens, or males are held; but it is nei-
ther easily induced nor is it continuous,
especially in males. Digging is induced
when nest soil is moistened. Refuse
heaps, including food wastes, discarded
cocoons, and dead immatures or adults,
are accumulated away from the occupied
sections of nests. Wastes are regularly
deposited in the small (15 mm in diame-
ter) dishes used for feeding honey, which
are frequently filled with soil, even by
colony-founding queens. These seem to
be behaviorally totipotent compared
with workers. Males are occasionally ob-

served riding for many minutes on alate
female nest mates, without attempting
copulation. All standard self-grooming
routines (7) are actively practiced, ex-
cept abdominal tip licking, which has not
been seen. Alarm communication is slow
and inefficient. The ants are generally
nonaggressive, differing markedly from
most Myrmecia. Territoriality between
colonies is not evidenced in either field
or laboratory. Allozyme markers suggest
that several colonies can contribute for-
agers to single trees. Workers trans-
ferred to alien colonies are shown little
aggression. Mixed foragers will settle to
behave like queenless colony fragments
and will adopt foraging queens.

Thus, almost all behavioral character-
istics of Nothomyrmecia are held in
common with Myrmecia, further con-
firming the primitiveness of Nothomyr-
mecia .

Retrospect

The resistance to collection for 46
years by Nothomyrmecia is explained by
its apparently patchy, locally limited dis-
tribution, its nonspecific, insignificant
nest entrances, its strict nocturnality of
foraging, the likely restriction or cessa-
tion of aboveground activity in winter
(correlating negatively with that of most
insect collectors in Australian .areas of
severe summer climate), and the unlikely
collection at light traps of the brachyp-
terous queens (males in this context
would probably not have been recog-
nized). ' Doubts have been expressed
about the true provenance of the original
specimens, which I believe were very
likely collected in mallee woodland
south of Balladonia. The previously al-
most exclusive emphasis by would-be
collectors in this area on heath rather
than mallee sites could have been mis-
directed. The rediscovery of Nothomyr-
mecia in Western Australia remains a
challenge to interested naturalists.

References and Notes

1. J. Clark, Mem. Nail. Mas. Victoria Melbourne
8, 17 (1934).

2. , The Formicidae of Australia, (I) Sub-
family Myrmeciinae (Commonwealth Scientific
and Industrial Research Organisation, Mel-
bourne, 1951), p. 16.

3. W. L. Brown, Insectes Soc. 1, 22 (1954).

4. and R. W. Taylor, in The Insects of Aus-
tralia (Melbourne Univ. Press, Carl ton. Victo-
ria, 1970), p. 958.

5. E. O. Wilson, F. M. Carpenter, W. L. Brown,
Science 157, 1038 (1967); Psyche 74, 1 (1967).

6. C. P. Haskins, in Development and Evolution of
Behavior, L. R. Aronson, E. Tobach, D. S. Lehr-
man, I. S. Rosenblatt, Eds. (Freeman, San
Francisco, 1970), p. 355.

7. E. O. Wilson, The Insect Societies (Harvard
Univ. Press, Cambridge, Mass., 1971).

8. W. L. Brown and E. O. Wilson, West. Aust.
Nat. 1, 25 (1959).

9. The party included D. H. CoDess, J. E. Feehan,
J. D. Lawrence, M. S. Upton, and R. W. Tay-

SCIENCE, VOL. 201

lor, all attached to the Australian National In-
sect Collection, CSIRO, Canberra. The first
Nothomyrmecia specimens were collected by
Taylor.

10. D. C. F. Rentz and M. J. D. White, personal
communication.

11. W. M. Wheeler, Schr. Physikal-Okonomischen
Ges. Konigsberg SS, 1 (1914).

12. M. J. Viana and J. A. Haedo Rossi. Ameghi-
niana 1, 108(1957).

13. G. S. Tulloch, Ann. Entomol. Soc. Am. 29, 81
(1936).

14. H. Markl, Proc. Congr. IVSSl 7th London
(1973), p. 258; terminology here follows that of
P. D. Ashlock and J. D. Lattin [Ann. Entomol.
Soc. Am. 56,693(1963)].

15. D. J. Brothers, Univ. Kans. Sci. Bull. 50, 483
(1975).

16. C. P. Haskins and E. F. Haskins. Insectes Soc.
2, 115(1955).

17. W. L. Brown and W. L. Nutting, Trans. Am.
Entomol. Soc. 75, 113 (1950); W. L. Brown.
Psyche 72, 65 (1965).

18. E. F. Riek, in The Insects of Australia (Mel-
bourne Univ. Press, Carlton, Victoria. 1970), p.
874; P. Dessart and L. Masner, Bull. Ann. Soc.
R. Entomol. Belg. 101, 275 (1965); L. Masner.

Can. Entomol. 108, 1243 (1976); and

P. Dessart, Bull. Inst. R. Sci. Nail. Belg. 43, 1
(1967).

19. J. Forbes, J. N. Y. Entomol. Soc. 75, 35 (1967).

20. H. R. Hermann, J. Ga. Entomol. Soc. 4, 123
(1969).

21. T. Eisner, Bull. Mus. Comp. Zool. Harv. Univ.
116,439(1957).

22. G. C. Wheeler and J. Wheeler, Am. Midi.
Nat. 48, 111 (1952); Pan Pacif. Entomol. 47,
245 (1971).

23. , Mem. Entomol. Soc. Wash. 7 (1976).

24. The areas identified here as acrosclerites might
in fact be secondarily differentiated sections of
the definitive tergite and stemite. This would not
substantially affect the argument. The term

"pretergital belt" was used for the anterior por-
tion of the segment IV exoskeleton by W. L.
Brown, Search Agric. Geneva N.Y. 15, 37
(1975).

25. G. C. Crampton, J. N.Y. Entomol. Soc. 39, 323
(1931).

26. Lateral fusion of segments III, IV, and V in ants
was briefly reviewed by W. H. Gottwald [Cor-
nell Univ. Agric. Exp. Sin. Mem. 408, 126 (1969)].
In species of Heteroponera, Gnamptogenys,
Proceratium, Discothyrea (all Ectatommini),
and Psalidomyrmex (Ponerini) abdominal IV is
strongly fused laterally and arched or reflexed,
turning the gastral apex ventrally or even ante-
riorly [W. L. Brown, Bull. Mus. Comp. Zool.
Harv. Univ. 118, 175 (1958)]. Asphinctopone
(Ponerinae) workers have segment IV tubulate,
with laterally fused sclerites. However, the
acrosclerites are (presumably through second-
ary reduction) mere ridges, with the insertion in-
to segment III being only slight. Sphinctomyr-
mex (Cerapachyini) has serially repeated tubula-
tion of segments IV, V, and VI, producing sev-
eral gastral constrictions.

27. Development of a strongly constricted post-
petiole has independently occurred within sev-
eral ponerine genera including Proceratium (Ec-
tatommini) and Cerapachys (Cerapachyini), and
in the evolution from primitive ponerine or po-
neroid stock of Dprylinae, Ecitoninae, Leptanil-
linae, and Myrmicinae. The only ecitonine ge-
nus usually said to lack a worker postpetiole is
Cheliomyrmex (Ecitoninae-Cheliomyrmecini); it
has a tubulate but relatively unconstricted seg-
ment IV, resembling that of many ponerines
(such as Amblyopone australis) (Fig. 3A) and
exemplifying a condition presumably primitive
for Ecitoninae [see figure 14-29 of (7), p. 70].
Similarly, in the Dorylinae, Dorylus has ab-
dominal structure like Cheliomyrmex , while Ae-
nictus workers have a strongly constn/cted post-
petiole.

28. This, perhaps once major, group includes

the living relict Aneurelus simoni (Sri Lanka)
and three extinct Oligocene genera of Europe
or North America [E. O. Wilson, T. Eisner,
G. C. Wheeler, J. Wheeler, Bull. Mus. Comp.
Zool. Harv. Univ. 11!>, 81 (1956)].

29. T. Eisner and W. L. Brown, Proc. Int. Cong.
Entomol. 10th 2. 503 (1958); T. Eisner and E. O.
Wilson, ibid., p. 509.

30. F. Invrea, Ann. Mus. Civ. Star. Nat. Genova
69, 257 (1957).

31. H. Kutter, Mitt. Schweiz. Entomol. Ges. 23, 81
(1950); Du 11 (4), 45 (1951). An illustration re-
produced as figure 19-1 in (7), p. 350.

32. T. C. Schneirla, Army Ants, a Study in Social
Organization (Freeman, San Francisco, 1971);
see also figure 4-22 in (7), p. 60.

33. A. D. Bishop and R. H. Crozier, personal com-
munication.

34. H. T. Imai, R. H. Crozier, R. W. Taylor,
Chromosoma 59, 341 (1977).

35. P. S. Ward, personal communication.

36. These include species of Santalum and Exo-
carpos (Santalaceae). Comesperma (Poly-
galaceae), Dianella (Liliaceae). Amyema (Loran-
thaceae), Enchylaena (Chenopodiaceae), Geij-
era (Rutaceae), Callitris (Cupressaceae), Eu-
calyptus and Melaleuca (Myrtaceae), as
identified by R. Pullen, I. Brooker, and G. Chip-
pendale (personal communication).

37. J. J. McAreavey, Proc. Linn. Soc. N.S.W. 73,
137 (1948).

38. P. J. Darlington, Ecol. Monogr. 13, 37 (1943).

39. W. M. Wheeler, Proc. Natl. Acad. Sci. U.S.A.
3, 109 (1917).

40. D. T. Briese and E. A. Davison, personal
communication.

41. Others, in addition to A. D. Bishop, R. H. Cro-
zier, and P. S. Ward (Australia), involved in the
research on Nothomyrmecia discussed in this
article are T. Eisner, B. Holldobler, G. C.
Wheeler, J. Wheeler, E. O. Wilson (United
States); P. Duelli (Switzerland); and H. Markl
(Germany).

SCIENCE. VOL. 201, 15 SEPTEMBER 1978

0036-8075/78/0915-0985$02.00/0 Copyright 1978 AAAS

985